In this paper the possibility of detecting and locating damages in a 20 m high steel lattice mast subjected to natural excitation has been investigated. For the damaged mast seven different damage states were considered. In these damage states a damage was assumed in one of the lower diagonals. These diagonals were cut and provided with a bolted joint implying that a damage could be simulated. Based on 20 periodical measurements during 6 months the sensitivity of the modal parameters, identified by an ARMA-model, to environmental conditions such as wind-direction, wind-speed and air-temperature have been investigated. These sensitivities have been compared with the changes of modal parameters due to a damage. It is found that the measured natural frequencies vary less than one per cent while the measured modal damping ratios vary more thai1 twenty per cent due to different environmental conditions. The measured bending natural frequencies and the measured rotational frequency approximately decrease few per cent and more than ten per cents, respectively, due to a damage corresponding to a removal of one of the lower diagonals. The results also show that a neural network trained with simulated data is capable for detecting location of a damage in the steel lattice mast when the network is subjected to the experimental data.
Proceedings of "dynamics of Structures": A Workshop on Dynamic Loads and Response of Structures and Soil Dynamics, September 14-15, 1994, Aalborg University, Denmark, 1994
System Identitication; ARMA Models; Damage Detection; Civil Engineering Applications; Neural Networks
Main Research Area:
Dynamics of Structures : a workshop on dynamic loads and response of structures and soil dynamics, 1994
Dept. of Building Technology and Structural Engineering, Aalborg University